Advertisement

Can someone explain why studios making large space games are congratulated when they make their game capable of using double precision, allowing it to be very large?

Started by September 11, 2024 05:00 AM
13 comments, last by frob 1 month, 3 weeks ago

Alberth said:
How large is large?

This is also the problem with the games.

It is rare for space games to actually use a large space. Although they may thematically have planets and outposts across the galaxy, the actual gameplay is quite small.

Space is HUGE. As already touched on in other replies, fast transportation is always used, with jump points or hyperspace or some other teleport method. If you set your starship to travel 10x the speed of light, the nearest star is still 4.24 light years away, so the journey will take about 5 months.

Traveling around a spaceport, that's generally a small model under 1 km virtual size.

Dogfighting, mining, and all the other ‘interesting’ stuff all take place in relatively small regions. The numbers on the display are there for emotional purposes, they aren't typically an actual 1:1 scale of the simulated distances.

Content at the scales is hard. The AAA games taking millions of man-hours to create and filled with content tend to be far smaller than people imagine. Playable areas tend to only be a few kilometers across in any direction. BotW was about 80 square kilometer. Skyrim's playable area was about 40 square kilometers. Ark maps are between 36 square km to 80 square km. GTA5 has an area of around 50 square km with enormous holes in it. Even though many games can feel like they represent enormous Earth-sized spaces or large sprawling cities, in reality even the biggest ones are smaller than a single suburban area.

Space games, actually encountering other players is hard. Take EVE Online when it was around 10M active players and 20K concurrent players, they'd still be mostly clustered in a few of the game's roughly 8000 zones, most of the zones were vacant for active humans at any given time.

Studios making large space games are congratulated for implementing double precision because it represents a significant technical achievement that overcomes a major limitation in game development. Implementing double precision represents a significant technical achievement that enables the creation of vastly larger and more realistic game worlds, particularly beneficial for space simulations and other large-scale games.

Advertisement

Tphone12 said:
double precision because it represents a significant technical achievement that overcomes a major limitation in game development

How do you figure that? There's nothing inherently hard about double precision. For code it's writing “double” instead of “float”. For C-derived languages double is the default type, A constant “1.234” is double and needs a modifier “1.234f” (the f at the end) to mark it as a float.

The hardware is slower if double is present at all, in which case it takes software emulation.

Tphone12 said:
enables the creation of vastly larger and more realistic game worlds

Example or citation needed. Counterexamples are plentiful and already given in earlier replies to the thread. Historically the largest and widely-played space games rely on single-precision floating point, with submaps and re-centering when needed. What exactly do you believe it enables that wasn't possible otherwise?

This topic is closed to new replies.

Advertisement